科技通报
科技通報
과기통보
BULLETIN OF SCIENCE AND TECHNOLOGY
2015年
6期
166-168
,共3页
模糊图像%光照%图像分割
模糊圖像%光照%圖像分割
모호도상%광조%도상분할
fuzzy image%illumination%image segmentation
对模糊图像的多尺度分割,是解决许多计算机视觉处理问题的基础。传统的图像分割算法采用基于小波变换的局部特征匹配方法,无法有效去除光照的干扰,对运动目标图像的分割效果不好。提出一种基于模糊图像边缘能量特征提取的运动目标图像的去光照干扰分割方法。计算去光照干扰后的运动目标图像振幅分量和频率分量,采用混合函数控制曲线方法生成运动目标图像时间序列,计算每个尺度下计算运动目标图像的边缘能量特征,进行图像区域特征的非同态块匹配分割,最终生成灰度直方图二进制均衡系数,实现了运动目标图像的准确分割,去除了光照干扰。仿真结果表明,该算法具有分割结果准确,抗干扰能力较好,图像分割质量较优。
對模糊圖像的多呎度分割,是解決許多計算機視覺處理問題的基礎。傳統的圖像分割算法採用基于小波變換的跼部特徵匹配方法,無法有效去除光照的榦擾,對運動目標圖像的分割效果不好。提齣一種基于模糊圖像邊緣能量特徵提取的運動目標圖像的去光照榦擾分割方法。計算去光照榦擾後的運動目標圖像振幅分量和頻率分量,採用混閤函數控製麯線方法生成運動目標圖像時間序列,計算每箇呎度下計算運動目標圖像的邊緣能量特徵,進行圖像區域特徵的非同態塊匹配分割,最終生成灰度直方圖二進製均衡繫數,實現瞭運動目標圖像的準確分割,去除瞭光照榦擾。倣真結果錶明,該算法具有分割結果準確,抗榦擾能力較好,圖像分割質量較優。
대모호도상적다척도분할,시해결허다계산궤시각처리문제적기출。전통적도상분할산법채용기우소파변환적국부특정필배방법,무법유효거제광조적간우,대운동목표도상적분할효과불호。제출일충기우모호도상변연능량특정제취적운동목표도상적거광조간우분할방법。계산거광조간우후적운동목표도상진폭분량화빈솔분량,채용혼합함수공제곡선방법생성운동목표도상시간서렬,계산매개척도하계산운동목표도상적변연능량특정,진행도상구역특정적비동태괴필배분할,최종생성회도직방도이진제균형계수,실현료운동목표도상적준학분할,거제료광조간우。방진결과표명,해산법구유분할결과준학,항간우능력교호,도상분할질량교우。
The segmentation of multi scale fuzzy image processing, computer vision is to solve many problems of the founda?tion. Traditional image segmentation algorithm using local feature matching method based on wavelet transform, can effec?tively remove the interference of light, it is not good for the movement of the target image segmentation effect. This paper presents a moving object image to extract the fuzzy image edge energy feature based on light interference to segmentation method. Calculation of interference of light after the moving target image amplitude and frequency components, using mixed function control curve generation method of moving target image time series, the edge energy characteristic calcula?tion of object image is calculated for each scale, feature of image region non homomorphic block matching segmentation, the final generation of gray histogram binary equilibrium coefficient, the realization of accurate segmentation of moving tar?get image, removed the interference of light. The simulation results show that, the algorithm has a good segmentation result, better anti-interference ability, the quality of image segmentation is better.